interesting behavior from rpctrace

[Known Human Nick Rusnov <nick@grawk.net>]

Something pretty interesting happens when I try to run apt-get from rpctrace:

fugue:~# rpctrace apt-get sounce clisp
task83-> 2030 () = 0 {4096 2387 1738 2478 584 5856 0 1880 317 46283 4515 3414 3239}
task83-> 2029 (134217728) = 0 134512640 118784 5 7 1 0   67 0
task83-> 3206 (pn{  5})rpctrace: ../../utils/rpctrace.c:522: trace_and_forward: Assertion `(((inp->msgh_bits) & 0x0000ff00) >> 8) == info->type' failed.
Aborted

But whats really interesting about this is what happens *after* it says Aborted; there is some
extensive grinding and ps -Av shows something interesting.

First, a snippet before the above command:

  PID STAT    SZ   RSS %CPU %MEM COMMAND
    2 D<p  56.2M 1.73M  0.0  5.4 /boot/gnumach.gz root=hd0s3

Then after:

    2 D<p  56.1M 10.1M  0.0 31.8 /boot/gnumach.gz root=hd0s3

Notice the huge jump in resident size. It doesn't go down over time.

Doing the same rpctrace several times in a row results in a slight increase in resident
storage. Done in rapid succession it causes a crash which spontaneously reboots
the computer.

as always,
nick
            nick@grawk.net * http://www.fargus.net/nick
    Developer - Systems Engineer - Mad System Guru - MOO Sales
    He picks up scraps of information/He's adept at adaptation
Because for strangers and arrangers/Constant change is here to stay